公开(公告)号：DE10050171A1

公开(公告)日：2001-04-19

申请号：DE10050171

申请日：2000-10-11

Applicant: IBM

Inventor： KANAMARU ATSUSHI ,  ASANO HIDEO ,  KIBASHI AKIRA ,  SAITO TAKAHIRO ,  KOBAYASHI KEIJI

IPC: G06F3/06 ,  G06F12/08 ,  G06F12/12

Abstract: Involves operating high order cache-memory and the low order cache-memory in different data 'exchange' modes. When a system is accelerated or run up to speed, an exchange of selected data for selecting a data 'exchange' mode for each cache-memory between the cache-memory of the high order and the cache-memory of the low order, takes place. Avoidance of errors in data storage in cache-memories used in computer systems based on hard-disc drives (HDDs) and with high and low order cache-memories being used and operated in different data 'exchange' modes, such as the least recently used (LRU) and least frequently used (LFU) modes, when a cache-memory has no free data storage locations available, involves operating the high order cache-memory and the low order cache-memory in different data 'exchange' modes, and when a system is accelerated or run up to speed, an exchange of selected data for selecting a data 'exchange' mode for each cache-memory between the cache-memory of the high order and the cache-memory of the low order, takes place..

公开(公告)号：DE10050171B4

公开(公告)日：2006-01-05

申请号：DE10050171

申请日：2000-10-11

Applicant: IBM

Inventor： KANAMARU ATSUSHI ,  ASANO HIDEO ,  KIBASHI AKIRA ,  SAITO TAKAHIRO ,  KOBAYASHI KEIJI

IPC: G06F3/06 ,  G06F12/08 ,  G06F12/12

Abstract: A system and method are provided that reduce the amount of data held commonly in both high-ranking and low-ranking cache memories, thereby having each of those cache memories hold data more efficiently. More particularly, a computer system is provided with an HDC card 21 connected to an expansion bus 20 and an HDD unit 22 connected to the HDC card 21. The HDC card 21 is provided with a disk cache (high-ranking cache memory) and the HDD unit 22 is provided with a disk cache 54 (low-ranking cache memory). The HDC card 21 and the HDD unit 22 exchange select information for selecting a swap mode of each cache memory when the system is started up, thereby selecting different swap modes according to the exchanged select information respectively.

公开(公告)号：JP2001216092A

公开(公告)日：2001-08-10

申请号：JP2000019379

申请日：2000-01-27

Applicant: IBM

Inventor： UCHIIKE HIROSHI ,  KURODA TAKASHI ,  KANAMARU ATSUSHI ,  NISHINOMIYA HIROMI ,  SAITO TAKAHIRO

IPC: G06F13/10 ,  G06F3/06 ,  G11B5/09

Abstract: PROBLEM TO BE SOLVED: To improve the reading efficiency to a write command that is kept waiting for its execution and also to improve performance of a hard disk drive. SOLUTION: When a data read request is received while a data write request is kept waiting for its execution, the logical block address of a write command W1 is compared with that of a read command R for deciding presence or absence of their duplication in a positional relation deciding step and a data reading step, where the read command R is executed before all write commands W are executed when no duplication of block addresses is decided.

公开(公告)号：JP2001117817A

公开(公告)日：2001-04-27

申请号：JP29390999

申请日：1999-10-15

Applicant: IBM

Inventor： KANAMARU ATSUSHI ,  ASANO HIDEO ,  KIHASHI AKIRA ,  SAITO TAKAHIRO ,  KOBAYASHI KEIJI

IPC: G06F3/06 ,  G06F12/08 ,  G06F12/12 ,  G11B19/02

Abstract: PROBLEM TO BE SOLVED: To efficiently hold data in respective cache memories by holding data which are common to a high-order and a low-order cache memory or decreasing the number of pieces of data held in the high-order and low-order cache memories in common. SOLUTION: A computer system is equipped with an HDC card 21 which is connected to an extension bus 20 and an HDD drive 22 which is connected to an HDC card 21. The HDC card 21 is equipped with a disk cache (high-order cache memory) and the HDD drive 22 is equipped with a disk cache 54 (low- order cache memory). The HDC card 21 and HDD drive 22 send and receive selection information for selecting swap modes of the cache memories when the system is started and selects mutually different swap modes according to the sent and received selection information.

公开(公告)号：JPH10134528A

公开(公告)日：1998-05-22

申请号：JP28383196

申请日：1996-10-25

Applicant: IBM

Inventor： OZEKI HIDENORI ,  KIBASHI AKIRA ,  NISHINOMIYA HIROMI ,  SAITO TAKAHIRO

IPC: G11B19/04 ,  G11B20/18

Abstract: PROBLEM TO BE SOLVED: To efficiently execute ERP that is executed to prevent an error when reading a signal on a disk by grouping steps registered in an error recovery procedure(ERP) and selecting and executing a specific step out of the groups according to an error type. SOLUTION: In an ERP and its execution order change system, a host I/F 401 communicates with a host system and reports to a task handler 402 when it receives a command from the host. The task handier 402 sends a read/write system command to a read/write system processing mechanism 405 according to the command type and sends other commands to an other command processing mechanism 403. When an error occurs during a series of processing, a read/write manager 421 notifies of the location where an error occurred to an ERP routine 423 to execute ERP.

公开(公告)号：JP2001125829A

公开(公告)日：2001-05-11

申请号：JP30660599

申请日：1999-10-28

Applicant: IBM ,  PROMISE TECHNOLOGY INC

Inventor： KANAMARU ATSUSHI ,  ASANO HIDEO ,  KIHASHI AKIRA ,  KUSHIDA KOICHI ,  SAITO TAKAHIRO ,  CHICHEN UU ,  KELVIN KAO

IPC: G06F12/08 ,  G06F3/06 ,  G06F9/32 ,  G06F9/34 ,  G06F9/38 ,  G06F13/38

Abstract: PROBLEM TO BE SOLVED: To execute detailed analysis and prediction by an outside controller in reponse to a true request from an application, and to issue a look-ahead request to an auxiliary storage device such as an HDD. SOLUTION: An HDC card 21, which stores data and is connected to an HDD device 22 having a cache memory, is provided with an access request trace part 61 for directly tracing a true access request from an application program to be executed by a host from the application program, a speculative request deciding part 62 for deciding a speculative request estimated in the future based on the traced true access request, and an HDC 31 for issuing the decided speculative request to the HDD device 22.

公开(公告)号：JP2003122631A

公开(公告)日：2003-04-25

申请号：JP2001316036

申请日：2001-10-12

Applicant: IBM

Inventor： KUSHIDA KOICHI ,  KANAMARU ATSUSHI ,  SAITO TAKAHIRO

IPC: G06F3/06 ,  G06F12/08 ,  G06F12/12

Abstract: PROBLEM TO BE SOLVED: To efficiently execute a reading request issued from a host computer when a writing request is cached in a cache memory. SOLUTION: When a reading request is transferred, whether or not a cache memory is put into a memory state that the cache memory can cache the reading request is judged (S103). When it is impossible, the oldest writing request and the other prescribed writing request are executed after the order of execution is optimized so that the reading request can be cached by the cache memory (S107, S113). And, the cached reading request is executed preferentially to the writing request remaining in the cache memory by an RBPW (S111). Therefore, it is possible to efficiently execute the reading request.

公开(公告)号：JPH10275110A

公开(公告)日：1998-10-13

申请号：JP7494997

申请日：1997-03-27

Applicant: IBM

Inventor： KANAMARU ATSUSHI ,  KAKIHARA TOSHIO ,  KIGAMI YUJI ,  SAITO TAKAHIRO

IPC: G06F3/06 ,  G06F12/08 ,  G11B19/02

Abstract: PROBLEM TO BE SOLVED: To provide the disk drive device and its control method which can make an additional read of media at need while minimizing a command overhead, and then improve the total performance. SOLUTION: The disk drive device 10 performs data transfer until final one block is left and expects the intervention of the local MPU 16 even when HIC 15 is able to perform all data transfer without the intervention of a local MPU 16 while a read command is issued and requested data hit data held in a cache memory, the local MPU 16 makes an indication for transferring the final one block once preparations for the end of the command are made, and if there are more than one block of transfer data left when the local MPU 16 instructs the HIC 15 to transfer the final one block, the HIC 15 performs all the data transfer without interrupting the data transfer at the final one block, thus completing the command.

公开(公告)号：JPH10134516A

公开(公告)日：1998-05-22

申请号：JP28386096

申请日：1996-10-25

Applicant: IBM

Inventor： SAITO TAKAHIRO ,  KIBASHI AKIRA ,  OKADA KENJI ,  OZEKI HIDENORI

IPC: G06F3/06 ,  G11B20/10 ,  G11B20/12 ,  G11B20/18

Abstract: PROBLEM TO BE SOLVED: To minimize the reduction in a data transfer rate by arranging a spare sector at a position in a spare region located in each zone and is away from the disk center from the position of data in which a hardware error has been generated at least by a specific angle or more from the position of hard error being generated. SOLUTION: A zone 0 on a disk includes at least one spare track arranged at a center part in advance and a data region is divided into a data region 1 and a data region 2 by a spare track. When the zone 0 has a defective sector D, a spare track where a spare sector S is located is sought, the data of the spare sector S are read, and a track located at the next sector of the defective sector D is sought, thus reallocating the data of the sector D at the spare sector S at nearly the opposite side for the disk center in the spare track.